►
Description
Join the GitHub Archive Program partners for a deep dive into their unique approaches and technologies used to preserve the world’s knowledge for future generations.
Featuring (in order of appearance):
- Nick Brysiewicz, Director of Development, The Long Now Foundation
- Patricia Alfheim, Communication Manager, Piql AS
- Bendik Bryde, Business Development Manager, Piql AS
- Ioan Stefanovici, Senior Researcher, Microsoft (Project Silica)
- Henry Lowood, Harold C. Hohbach Curator for History of Science & Technology Collections, Stanford University
- Roberto Di Cosmo, Founder, Software Heritage
- Brewster Kahle, Founder & Digital Librarian, the Internet Archive
About GitHub Universe:
GitHub Universe is a two-day conference dedicated to the creativity and curiosity of the largest software community in the world. Sessions cover topics from team culture to open source software across industries and technologies.
For more information on GitHub Universe, check the website:
https://githubuniverse.com
A
Good
afternoon,
everyone
and
welcome
to
this
next
set
of
talks
about
the
state
of
the
art
of
long-term
thinking
and
of
long-term
archiving.
My
name
is
nicholas
paul
breisowitz,
I'm
the
director
of
development
at
the
long
now
foundation,
and
I
currently
have
the
curious
task
of
condensing
the
next
and
last
10
000
years
into
the
next
eight
minutes.
So
please
bear
with
me
if
it
seems
like
I'm
talking
a
little
quickly.
I've
got
a
lot
of
cool
stuff.
A
I
want
to
walk
you
guys
through
and
I'm
excited
to
get
through
all
of
it
today.
So,
let's
see,
if
we've
got
this
perfect,
so
the
long
now
foundation
is
a
non-profit
here
in
san
francisco,
california,
that
was
founded
in
1996,
with
the
hope
of
getting
people
to
think
more
seriously
about
the
long
term.
So
for
us
in
this
case
this
is
the
next
and
last
10
000
years
of
human
civilization.
A
So,
as
opposed
to
the
short
now
of
this
conference
or
of
this
semester,
this
election
cycle,
this
business
quarter
we're
trying
to
get
people
to
think
about
this
20
thousand
year.
Narrative
that
we
call
the
long
now
and
one
of
the
advantages
of
this
framework
is
that
putting
yourself
in
the
middle
of
a
twenty
thousand
year.
Temporal
window
like
this
means
you're
the
inheritors
of
ten
thousand
years
of
art,
culture,
science
and
technology.
A
But
it
also
means
that
you're,
in
the
middle
of
the
story
and
at
the
beginning
of
the
next
10
000
years,
at
an
extraordinarily
leveraged
point,
where
you
have
a
lot
of
responsibility
to
be
a
good
ancestor
to
the
next
400
generations
of
humans
that
are
coming
after
you.
But
of
course
most
people
aren't
really
thinking
that
far
out
and
that's
where
this
guy
comes
in.
A
This
is
danny
hillis,
who
some
of
you
might
know
as
one
of
the
people
who
was
responsible
for
inventing
massive
parallel
computing
and,
like
so
many
technologists
of
his
age,
danny
realized
in
the
90s
that
the
year
2000,
where
they
had
always
thought
of
the
future
being
being
kind
of
out
there
by
the
year.
2000.
That's
when
all
this
stuff
was
going
to
happen.
A
It
was
going
to
come
and
go
in
the
blink
of
an
eye,
and
we
didn't
really
have
a
replacement
symbol
for
the
future,
something
to
get
people
to
inspire
people
to
think
long
term
and
so
danny
being
something
of
a
computer.
Geek
thought
hey.
I
should
make
a
mechanical
clock.
The
size
of
a
mountain
that'll
keep
good
time
for
the
next
10
000
years
and
will
act
as
an
icon
and
a
symbol
for
long-term
thinking
in
the
future.
A
A
That
would
ensure
that
every
visitor
to
the
clock
on
every
single
day
in
the
next
10
000
years
would
hear
a
bell
chime
that
no
human
being
had
ever
heard
before
and
no
human
being
would
ever
hear
against.
That's
about
3.65
million
unique
bell,
chime
sequences
that
are
programmed
into
the
clock
and
that
clock
is
now
currently
under
construction.
Inside
of
this
mountain
in
west
texas,
they've
been
building
this
thing
for
most
of
my
adult
life
and
it's
still
under
construction
right
now.
A
Rumor
has,
it
might
be
done
within
my
lifetime,
but
we've
never
actually
offered
up
a
completion
date
for
it.
But
I
did
bring
a
really
cool
video
to
share
with
you
really
quickly
about
some
of
the
installation
work
that
went
on
with
the
bell,
the
chime,
winder
or
sorry,
the
the
belgium
generator
and
the
clock
winder
last
year,
and
it
does
have
some
sound.
A
So
the
clock
is
currently
under
construction,
but
that's
not
the
only
way
we
can
get
people
to
think
long
term
for
some
people
the
clock
resonates
and
for
other
people
there's
other
ways
we
should
be
reaching
them,
and
so
we
have
a
host
of
other
projects.
One
of
the
other
personalities
that
got
involved
with
the
long
now
foundation
early
on
is
stuart
brand.
Here,
a
counterculture
figure
american
author
started
the
well
published
the
whole
earth.
A
Catalog
steward
thought
that
if
you're
gonna
go
to
the
trouble
to
bury
a
clock
for
somebody
to
find
in
a
few
thousand
years,
you
might
want
to
put
something
with
the
clock
that
would
get
them
inspired
about
disc
this
discovery.
So
what
kind
of
cultural
things
could
you
put
there?
Mostly?
He
was
thinking
about
a
library.
He
was
thinking
of
something
where
we'd
have
all
the
books
that
you
would
need
to
jump,
start
civilization,
so
specifications
for
how
to
make
good
steel,
socratic
dialogue,
shakespearean,
sonnets
the
whole
works.
A
You
want
to
put
this
together
and
have
it
there
for
somebody
to
find
with
the
clock.
But
of
course
you
get
into
these
language
questions
really
quickly
when
you
start
thinking
about
a
project
like
this,
which
is
you
know,
do
you
put
boar
haze
in
the
library
in
spanish?
Do
you
put
dostoyevsky
in
the
library
in
russian
or
do
you
just
put
the
whole
thing
in
english
or
the
whole
thing
in
mandarin?
A
Of
course,
this
is
an
homage
to
the
original
rosetta
stone
and
each
page
is
about
five
human
hairs
wide.
So,
rather
than
actually
encoding
data,
digitally
we've
actually
just
have
a
human
eye,
readable
format,
it's
very
small,
but
a
jeweler's
loop
will
get
you
there
about
a
hundred
times
magnification.
You
can
read
the
whole
thing
in
2016,
with
the
help
of
the
european
space
agent
esa.
We
were
able
to
land
one
of
these
archives
on
comet
67p,
and
then
we,
the
summer
we
landed
one
on
the
lunar
surface
as
well.
A
We
also
have
a
bunch
all
around
the
planet
because
in
archiving
there's
this
idea
called
locks.
L-O-C-K-S-S
which
stands
for
lots
of
copies,
keep
stuff
safe,
and
so
we
have
about
a
hundred
of
these
all
around
the
globe.
And
of
course,
if
you're
going
all
the
work
to
protect
languages,
the
question
is
well:
could
you
protect
and
preserve
species?
Could
you
use
current
genomic
technology
to
de-extinct
species
that
have
gone
extinct
like
the
woolly
mammoth
or
the
passenger
pigeon?
So
we
have
the
revive
and
restore
project
which
is
working
on
that.
A
A
Of
course,
there's
a
membership
group
around
long
now
he
has
about
11
000
members
across
60
countries,
and
then
we
decided
we
needed
a
physical
place
to
congregate,
so
we
turned
our
headquarters
and
fort
mason
center
just
down
the
street
into
the
interval.
Where
we
have
talks,
we
have
some
prototype
parts
of
the
clock.
Of
course
we
have
that
library
I
was
talking
about
and
it's
also
an
award-winning
cocktail
bar.
A
We
regularly
grace
a
number
of
top
ten
lists
in
san
francisco
and
across
the
united
states,
and
the
idea
being
that
you
would
go
with
a
friend
go,
have
a
cocktail
in
this
unusual
bar
surrounded
by
unusual
objects
and
you'd.
Ask
yourself
the
really
important
questions
from
being
a
good
ancestor
about
the
next
10
000
years
like
how
can
we
be
better
ancestors?
Of
course,
you
know
on
long
time,
skills
anticipating
future
need
becomes
very
difficult.
A
B
Good
afternoon
everyone,
so
we
are
pickle
from
norway,
we're
also
the
founders
behind
the
arctic
world
archive
that
you
saw
this
morning.
So
today
we
are
going
to
talk
about
our
approach
to
the
challenge
of
of
taking
care
of
information
into
the
perpetuity
and
how
we
use
simple
means
of
technology
to
ensure
the
access
of
that
data.
B
So
what
you
see
here
this
is
our
claim
to
fame.
This
is
our
core
competence.
It's
a
digital
time
capsule.
We
call
it
the
pickle
fill.
So
it's
an
offline,
storage,
medium.
It's
what
we
call
write
one's
read
many
medium,
so
it's
information
cannot
be
hacked.
The
information
cannot
be
tampered
or
taken
away
changed
on
that
storage
medium.
B
When
we
started
out,
we
wanted
to
design
a
storage
medium
that
could
last
for
centuries
and
even
thousands
of
years
and
coming
from
the
film
industry
we
saw,
there
was
some
properties
with
the
film
knowing
that
the
history
of
film
is
more
than
100
years
and
had
very
good
characteristics
on
how
to
keep
information
for
the
long
term,
and
so
what
we
did
is
to
do
that
with
digital
content
and
to
be
able
to
put
that
on
the
robust
medium.
So
there's
a
few
important
aspects
so
in
terms
of
authenticity.
B
So
in
the
digital
world
today,
with
deep
fakes
and
similar
tech
kind
of
emerging
technologies,
how
can
we
really
ensure
authenticity
in
a
thousand
year
perspective?
So,
from
our
point
of
view,
we
think
the
best
way
would
be
to
keep
the
the
true
original
of
that
of
that
file
and
keep
it
on
a
very
long
term.
Storage,
medium,
not
migrating
that
information,
not
changing
it
from
formative
format.
B
So
we
are
putting
all
the
information
about
the
decoding
software
about
the
file
format.
Specifications
about
everything
you
need
to
be
able
to
retrieve
the
information
in
the
future
in
both
a
machine,
readable
language,
but
also
in
a
human,
readable
language,
so
kind
of
gives
the
future.
You
users
the
opportunity
without
us,
knowing
kind
of
the
the
knowledge
that
we
have
in
the
future.
The
software
or
the
computing
platforms
still
the
opportunity
to
have
several
options
on
how
to
read
back
the
information,
so
it
could
be
like
a
low
tech
visual
proxy.
B
You
can
even
we
even
include
the
file
form
specification
to
help
that
process
if
the
compiling
of
the
source
code
is
difficult
in
the
new
system,
and
we
are
now
also
including
a
virtual
machine
to
help
the
process
of
the
recoding
and
rendering
of
the
files
for
the
most
used
common
file
formats
for
preservation
today
to
make
that
retrieval
process
faster
and
more
cost
efficient.
So
what
you
see
here
this
is
the
the
storage
medium.
B
So
you
see,
there's
a
visual
frame
on
top
and
there's
the
digital
frame
at
the
bottom,
so
we're
using
an
optical,
readable,
2d
barcode.
So
we
have
developed
a
very
flexible
barcode
that
we
called
the
boxing,
the
boxing
library.
So
this
is
written
in
the
c
language.
It's
open
source
and
published
in
our
repo
on
github.
B
So
the
idea
of
using
the
c
language
was
that
we
it's
been
used
for
a
long
time.
It's
quite
simple
and
very
fast.
It
also
doesn't
have
a
lot
of
dependencies
around
it
at
the
same
time
to
be
able
to
kind
of
so
we
are
dispersing
the
files
through
a
lot
of
these
frames,
but
able
to
kind
of
put
that
together
and
organize
it.
B
We
also
developed
something
in
we
call
the
archival
file
system,
so
the
archival
file
system
is
also
open
source
written
in
the
c
language
and
it's
and
it
has
a
table
of
content
which
makes
it
possible
to
understand
both
the
the
visual
file
and
the
digital
file,
and
it
understands
the
file
format
and
can
point
directly
to
that
specific
file.
Format,
decoder
which
is
stored
on
the
medium
for
the
retrieval,
at
the
same
time,
we're
storing
this
table
of
content,
both
in
a
human,
readable
and
a
machine,
readable
format.
B
So
we
think
this
combination
of
the
visual
and
the
digital
really
creates
an
extreme
resilience
to
the
long
term
of
the
storage.
What
you
see
here
is
so
we're
also
including
metadata
representation
representation
information
in
both
formats.
So
what
you
see
here
is
what
we
put
on
the
on
the
deposit
for
the
github
in
this
case.
That
could
be
a
mission
statement.
It's
a
file
form
specification
of
deflate
and
gzip,
which
was
used
in
this
deposit
to
store
the
source
code
on
our
storage,
medium
in
the
arctic
world
archive
and
then
so.
B
How
can
we
retrieve
the
data?
So
today
you
can
use
a
data
scanner.
So
what
you
see
here
on
the
on
the
right
side
is
the
pickle
reader,
so
this
scans
the
frames
and
then
convert
it
back
to
the
original
file.
But
we
also
wanted
to
create
the
opportunity
to
get
the
information
back
with
just
the
most
simple
means.
B
So
by
us,
as
we've
seen
kind
of
including
the
information
about
the
decoding
software,
the
archival
file
system
and
the
file
form
specifications,
you
are
able
to
get
back
the
information
just
by
using
a
magnifying
glass,
some
kind
of
a
capturing
device
like
a
camera
and
a
computer
in
the
future
to
retrieve
that
information.
And
if
you
want
to
understand
more
about
the
software
tools
that
are
using,
please
check
out
our
repo
on
the
on
the
github
site.
So
now
I'm
just
giving
it
over
to
patricia.
C
Thanks
mandic,
so
you
saw
in
the
video
this
morning
the
github
arctic
code,
vaults,
is
set
inside
our
vault,
which
is
called
the
arctic
world
archive
it's
located
up
on
the
remote
norwegian
arctic
island
of
svalbard,
and
it
was
opened
in
2017.,
so
quite
young,
the
vault
itself
is
inside
the
decommissioned
mine.
You
saw
that
on
the
video
this
morning
and
the
region
is
stimula
is
a
demilitarized
area
regulated
by
an
international
treaty,
making
it
a
super
safe
and
politically
stable
area.
C
So
the
vault
itself
is
cold,
dark
dry,
low
on
oxygen
and
which
is
great
because
that
means
no
electricity
to
store
this
data
for
perpetuity
for
hundreds
of
or
thousands
of
years
making
it
super
sustainable,
which
is
really
important
and
also
actually
extends
the
life
of
the
film
as
well,
and
to
be
able
to
accurately
predict
how
this
medium
will
survive
over
time.
We've
done
a
number
of
different
testing
on
it,
so
the
key
one
being
accelerated
aging,
which
uses
temperature
to
simulate
the
aging
process.
C
It's
an
iso
standardized
process
and
it
takes
a
really
long
time,
so
we're
seven
to
eight
years
into
the
testing
process
of
this
and
we've
passed
the
750
year
mark
for
read
back
if
it
was
stored
at
room
temperature,
so
that
dramatically
increases
the
colder
and
the
drier
the
conditions.
So
we
can
say
very
confidently
that
in
ava
the
data's
going
to
survive
for
a
thousand
years
at
least
2000
years,
we're
going
to
continue
testing,
see
how
far
we
get.
But
we've
proven
read
back
so
far.
C
We've
also
conducted
numerous
stress
tests,
both
from
the
medium
and
individual
frames,
to
see
how
it
behaves
after
lots
and
lots
of
reading
how
it
handles
wear.
Scratches
we've
even
exposed
it
to
quite
a
lot
of
radiation,
just
to
see
how
it
went,
and
it's
proven
to
be
an
incredibly
robust
and
resilient
medium,
which
is
great
because
then
we've
got
a
very
safe
and
secure
place
in
the
arctic
circle
and
we've
also
got
a
really
robust,
medium,
perfect
combination.
C
So
since
the
arctic
world
archive
was
opened
in
2017,
we've
had
more
than
30
global
deposits
with
more
happening
all
the
time.
Quite
a
lot
of
institutions
want
to
make
deposits
of
their
valuable
data
and
safeguard
their
piece
of
world
memory.
While
we
build
a
nice
picture
of
the
whole,
so
this
is
growing
all
the
time.
C
A
D
So
last
week
at
ignite,
satya
talked
about
our
collaboration
with
warner
brothers,
whereby
we
stored
the
classic
1978
film
superman
inside
a
piece
of
glass
and
our
partnership
with
warner
brothers
really
reinforced
to
us
the
importance
of
the
so-called
third
copy
of
your
data.
So
the
third
copy
is
essentially
a
copy
of
your
data
with
the
property
that,
if
all
else
fails,
you
want
to
have
the
guarantee
that
you
can
always
go
back
and
re-read
and
read
your
the
third
copy
and
be
able
to
safely
recover
your
data.
D
This
morning,
matt
friedman
talked
about
the
importance
of
preserving
source
code
right,
think
about
the
millions
of
people
years
worth
of
developer
time
and
effort,
that's
gone
into
building
and
debugging
the
systems
that
we
have
today
right.
It's
a
legacy
that
we
want
to
be
able
to
preserve
for
future
generations
right
as
well,
as
you
know,
wikipedia
the
internet
archive
effectively
the
sum
total
of
the
world's
knowledge.
D
Now
the
bad
news
is
that
existing
storage
technologies
are
inadequate
for
this.
The
vast
majority
of
data
today
is
stored
on
hard
disk
drives
and
tapes
which
are
magnetic
media,
and
the
problem
is
that
over
time,
the
media
degrades,
and
so
you
get
failure
modes
like
bit
rot
or
in
the
case
of
optical
discs,
media
rotten.
This
is
example
of
what
that
actually
looks
like,
so
this
is
a
laser
disc
from
1986
that
was
storing
the
bbc
doomsday
project,
which
was
a
census
of
the
uk
with
a
bunch
of
other
information.
D
What
we
aimed
to
do
is
to
create
a
storage
technology
that
has
that
that
attempt
solves
these
problems
and
gives
you
the
guarantee
that
you're
going
to
be
able
to
store
your
data
safely
for
thousands
of
years,
and
then
we
want
to
leverage
the
economies
of
scale
of
the
cloud
to
be
able
to
offer
this
as
a
service
to
businesses
and
consumers
and
the
media,
for
this
is
glass
and
I'm
not
talking
about
any
sort
of
exotic
type
of
glass.
This
is
a
this.
D
So
this
is
a
worm
media
right
once
read
many
and
it's
very
persistent.
So
what
I
mean
by
that
is
that,
in
order
to
write
data,
what
we
do
is
we
make
permanent
changes
to
the
physical
structure
of
the
media,
and
so
what
that
means
is
that
once
data
is
written
in
it,
there
is,
it
does
not
degrade
or
it
doesn't
change
physical
structure
in
any
way.
D
So
there's
no
bit
rot,
there's
no
media
decay,
which
means
that
we
don't
need
to
scrub
the
media
and
of
course
the
consequence
of
all
of
these
properties
is
that
we
can
now
afford
to
leave
the
data
in
situ.
So
once
written,
we
don't
have
to
keep
moving
it
over
every
few
years
to
a
new
device,
which
means
that
we
don't
have
an
accumulated
energy
and
financial
cost,
and
in
addition
to
that,
we
are
less
prone
to
suffering
from
data
loss
while
copying
due
to
hardware
or
software
bugs.
D
So
I
want
to
share
with
you
a
quick
video
to
kind
of
show
you
just
how
resilient
of
a
media
glass
is
and
kind
of
give
you
a
sense
for
why
it's
important
to
understand
these
things
from
an
operational
perspective
in
terms
of
how
you
might
use
this
in
a
data
center,
so
water
and
humidity
has
traditionally
been
a
very
big
problem
for
archival
media
in
particular.
Ferromagnetic
media,
as
you
can
see
here,
glass
is
waterproof.
D
We
can.
It
is
resistant
to
electromagnetic
to
magnetic
radiation,
which
means
that,
in
the
event
of
things
like
solar,
flares
or
solar
storms,
your
data
is
going
to
be
safe
because
we
store
the
data
inside
the
bulk
of
the
glass
material.
That
means
that
surface
level
scratches
or
physical
interactions
are
not
really
a
problem
for
us
and
you
can
even
bake
it
and
you
can
even
use
a
flamethrower
on
it,
but
we
didn't
have
time
to
be
able
to
include
that
in
this
video.
D
The
the
the
takeaway
there,
though,
is
that
you
know
if
you
have
a
fire
in
your
data
center,
for
example
you're
you
can
put
out
the
fire
with
water,
even
in
the
worst
case.
You
know
you
can
always
replace
your
electronics,
but
at
least
you
have
the
peace
of
mind
and
the
guarantee
that
your
data
is
going
to
be
safe.
D
So
how
does
this
actually
work
well?
In
order
to
write,
we
use
a
femtosecond
laser,
so
this
is
a
pulsed
laser.
The
pulses
are
on
the
order
of
hundreds
of
femtoseconds,
so
that's
10
to
the
minus
15
seconds
just
for
reference
and
what
we
do
is
we
focus
the
laser
beam
inside
the
bulk
of
the
glass
material
and
at
the
point
of
focus,
we
get
these
little
voxel
nano
structures
that
are
made
of
these
parallel
nanoplanes
and
by
changing
the
amount
of
energy
that
we
put
in.
D
D
So
to
read
it
you
use
a
totally
different
technology,
you
don't
use
a
laser,
you
use
polarization
microscopy,
and
so
the
light
source
for
that
is
a
simple
led
and
the
read
head
effectively
has
optical
elements
to
focus
the
light
to
focus
the
light
on
the
desired
sector
and
polarization
components.
So
now
we
get
to
how
do
voxels
actually
store
data.
So
when
polarized
light
passes
through
a
voxel,
the
voxel
will
change
its
polarization
across
two
dimensions.
D
One
is
phase
and
the
phase
is
changed
based
on
how
deep
or
how
how
strong
the
voxel
modification
is,
and
the
second
is
angle
and
the
polarization
angle
will
be
changed
based
on
the
physical
orientation
of
the
voxel
in
the
glass
material.
So
those
are
the
two
dimensions
that
I
mentioned
during
the
write
process
that
you
can
modulate
and
so
effectively
in
order
to
read
the
data
back.
D
What
you
do
is
you
focus
on
the
layer
that
you're
interested
in
you
shine
light
with
a
known
polarization
into
the
medium
and
then
you
effectively
measure
what
is
the
change
in
that
polarization
that
each
voxel
imparted
on
the
light
and
to
do
that?
We
use
machine
learning
techniques,
state-of-the-art
computer
vision
techniques
and
in
the
picture
that
you
see
there,
that's
an
actual
picture
of
data
written
read
from
one
of
our
samples
and
we
use
machine
learning
to
extract
the
symbol,
values
that
were
encoded.
D
Of
course,
we
also
use
error,
correcting
codes
to
fix
any
residual
errors
from
the
raw
reading
process
and
using
these
techniques.
I'm
very
happy
to
say
that
in
the
past,
in
less
than
three
years
since
the
project
has
been
going,
we
are
now
the
densest
optical
storage
technology
in
the
world.
So
with
that,
I'm
going
to
wrap
up,
you
know
have
a
look
at
our
website.
D
If
you
want
more
details
on
the
project,
if
you
want
to
read
a
paper
about
it-
and
you
know,
if
you
want
to
work
on
the
coolest
storage
project
out
there
right
now
join
us,
we
have
a
number
of
career
opportunities.
We
have
a
number
of
sde
roles,
postdoc
researchers,
internships
so
come
talk
to
me
or
send
me
an
email
so
yeah.
Thank
you.
E
E
This
is
all
good,
but
in
the
immortal
words
of
james
stockdale,
who
am
I-
and
why
am
I
here?
For
starters,
I've
been
engaged
in
software,
history
and
preservation
for
a
long
time.
Here's
the
invitation
I
received
almost
exactly
30
years
ago
to
attend
the
first
conference
on
software
preservation
held
in
early
1990.
E
E
E
E
So,
back
to
the
question,
why
me
well,
I
work
on
software
heritage,
both
as
a
curator
trying
to
save
it
and
as
a
historian,
writing
about
it
next
question:
why?
Github,
a
few
years
ago,
I
worked
with
a
team
at
uc
santa
cruz
to
develop
a
unified
approach
to
preserving
what
we
call
cultural
software
objects
and
their
development
histories.
E
E
What
is
documented
by
software
and
github?
Besides
the
obvious
answer
of
software
development
history?
Let's
consider
some
other
examples.
The
virtual
apollo
guidance
computer
project
supports
computer
simulations
of
the
onboard
guidance
computers
used
in
the
apollo
programs.
Lunar
missions
clearly
important
for
the
history
of
technology
generally.
E
The
code
repository
for
this
project
is
on
github,
but
a
full
accounting
of
the
project
includes
other
documentation
such
as
issues
posts
or,
of
course,
the
project
website.
Researchers
need
access
to
context
and
documentation
found
outside
github
to
fully
understand
the
code
deposited
inside
github
business.
History
is
another
area
in
which
github-based
software
has
played
a
significant
role.
Bitcoin
and
airbnb
provide
two
examples
for
business
related
projects.
We
again
depend
on
access
to
external
resources
to
better
understand
these
projects,
and
we
may
run
across
restrictions
that
limit
access
to
code
and
documentation.
E
Documentation
is
also
deposited
on
github,
for
example,
the
us
national
archives
and
records
administration
nara
put
its
digital
preservation
framework
on
github
for
public
comment.
This
framework
presents
preservation
action
plans
for
a
variety
of
digital
record
and
media
formats,
such
as
email
or
audio
files.
E
And
finally,
an
area
close
to
my
own
heart,
entertainment,
notably
computer
games,
free
doom,
is
particularly
interesting
for
my
own
research
on
player
created
content
based
on
source
code
released
by
id
software.
It
lets
players
run.
The
huge
catalog
of
mods,
replays
and
speed
runs
created
over
the
25
his
25
year,
history
of
doom.
E
E
Well?
In
a
nutshell,
stanford
will
produce
curated
collections
documenting
selected
github
projects
in
depth.
These
collections
will
include
code,
but
only
begin
there.
Code
alone
is
not
enough.
We
will
dive
in
beyond
the
code
repository
to
harvest,
contextual
documentation
that
historians
and
others
will
use
to
make
sense
of
these
projects
and
their
impacts.
E
E
The
next
few
slides
introduce
the
I
t
environment
that
will
support
this
project.
We
start
with
two
robust
preservation
systems:
sdr
the
stanford
digital
repository
and
locks.
You
already
know
lots
of
copies
keep
stuff
safe.
Both
of
these
were
developed
at
stanford,
we're
investigating
strategies
for
capturing
digital
objects
for
github
project
collections
and
may
end
up
using
a
mix
of
methods.
E
E
E
Sdr
assets
are
discoverable
by
stanford's
online
catalog
called
search
works
which
is
available
on
the
open
web
and
exposed
to
the
usual
search
engines,
documents
and
software
assets
are
accessible
via
search
works
by
the
way
stanford
was
a
project
partner
in
preserving
virtual
worlds,
a
pioneering
project
sponsored
by
the
library
of
congress.
About
a
decade
ago,
this
project
produced
a
few
collections
around
historical
game
software,
such
as
the
dame
the
doom
collection
that
you
see
here.
E
E
Finally,
an
end
product
for
this
collaboration
could
be
online
library.
Exhibits
that
interpret
github
projects
spotlight,
which
you
see
here,
is
stanford's
application
for
showcasing
digital
content
in
easy-to-produce
exhibits,
metadata
added
to
the
digital
content.
We
ingest
from
github
and
other
sources
will
activate,
search
and
browse
features
in
spotlight
to
help
researchers,
students
and
visitors
to
interpret
the
history
of
these
projects.
E
F
So
hi
everybody,
I'm
very
happy
to
be
here,
I'm
a
long
time.
Computer
science,
researcher
hacker,
developer,
open
source
advocate
and
today
here
and
hear
me
my
head
of
director
of
social
heritage,
to
tell
you
about
what
we
are
going
to
do
together
with
guitar.
But
let
me
start
from
the
first
point:
when
we
were
all
talking
about
software.
Software
is
fundamental
in
our
world.
It's
powers,
our
industry
fuels
innovation.
It's
a
pillar
of
modern
research
is
an
essential
mediator
for
access
to
digital
information.
F
F
When
I
was
a
student
many
years
ago
too
many
years
ago,
I
was
using
a
book
by
harold
davidson,
the
structural
interpretation
of
computer
programs,
where
you
find
this
phrase,
programs
should
be
written
for
people
to
read
and
only
access,
sorry
for
machines
to
execute.
Well.
Of
course,
you
can
say
this
is
a
professional
telling
their
students.
F
If
they
turn
in
homework,
they
cannot
be
read,
then
they
will
get
a
bad
grade,
but
it's
more
deep
than
this
I
mean
the
person
that
can
read
the
piece
of
software
should
be
able
to
read
a
piece
of
software.
Maybe
you
next
week
when
you
need
to
fix
it
when
he
was
writing
this
many
years
ago
there
was
not
much
source
code
available
to
showcase.
F
Today,
30
years
after
the
beginning
of
this
free
software
movement,
we
have
an
incredible
amount
of
software
available
here,
for
example,
you
get
an
accept
of
the
source
code
of
the
polar
guidance
computer
and
you
see,
even
if
you
are
not
yet
familiar
with
the
assembler,
that
I'm
showing
here,
you
can
read
the
comments
and
you
see
that
these
people
were
actually
very,
very,
very
careful
in
programming.
Actually,
this
is
a
part
of
the
mentor
that
is
asking
the
astronaut
to
turn
around
the
lunar
landing
module,
and
you
see
the
programming
is
very
careful.
F
When
you
get
the
message
and
say
proceed,
I
mean
the
astronaut
push
on
the
button
say:
okay,
I
have
turned
the
ram
around
well,
the
programmer
just
checks.
If
he's
lying,
I
mean
you
jump
back
and
check
with
them,
tennis
writing
in
the
right
position,
but
then
there
are
other
jewels.
We
were
heard
about
the
doom,
but
you
remember
john
carmack's,
incredible
developer.
He
also
wrote
quake
3..
F
If
you
remember
one
of
the
ancestors
of
a
call
of
duty,
which
are
too
many
people,
including
my
kids,
were
spent
too
much
time
on,
and
this
is
an
incredible
routine,
which
was
embedded
in
the
source
code
that
allows
you
to
compute
one
over
the
square
root
of
x
in
a
particularly
sophisticated
way
by
avoiding
all
floating
point
operations.
I
will
not
read
the
comments,
because
some
of
the
comments
are
forbidden
to
less
than
18
years
old,
but
I
mean
you
can
just
check
it
out
offline.
F
My
slides
are
online
by
the
way,
as
elena
schuster
wrote
in
a
beautiful
piece
of
article
some
years
ago.
Actually,
source
code
is
an
open
window
in
the
mind
of
the
developers,
and
so
it's
an
important
asset
and
we
need
to
keep
it
because,
even
if
you
are
not
able
to
run
the
binaries,
we
can
still
read
it
and
learn
things
by
reading
it.
F
Okay,
but
unfortunately,
this
source
code
is
a
precious
asset.
We
otherwise
will
not
be
here,
but
it's
also
an
endangered
heritage.
If
you
look
at
the
past,
there
are
many
key
people
who
developed
the
technology
we
are
using
today
who
are
passing
away
and
we
are
losing
their
knowledge.
At
the
same
time,
if
you
look
at
the
future,
software
development
is
skyrocketing.
F
Actually,
looking
inside
our
archive,
we
can
tell
you
that
the
amount
of
original
content
in
developed
around
the
world
doubles
every
two
years,
and
yet
we
are
missing
key
elements.
Today
we
are
missing
a
global
catalogue.
I
fully
agree.
There
is
a
lot
of
interesting
software
on
guitar,
but
there
is
also
a
lot
of
interesting
stuff
which
is
not
around
guitar
and
software
moves
around
they
change
names
and
projects
starts
and
finish
and
disappear,
and
then
we
miss
archives.
I
mean
you
remember
who
remember
google
code,
anybody
remembers
it.
F
I
mean
it
was
shut
down:
2015.
1
million
500
000
projects
in
dungeon,
and
many
other
things
like
this
are
happening.
We
do
not
have
an
archive
for
software
and,
finally,
what
we
are
missing
is
a
research
infrastructure.
One
of
my
best
friends
is
directing
the
james
webb
space
telescope
program.
Here
in
30
years,
10
billion
dollars
spent
to
actually
look
at
the
origin
of
the
universe,
something
that
I
definitely
love.
F
How
comes
as
computer
scientists
technologies,
we
are
not
have
not
been
able
to
put
together
a
common
open
infrastructure
for
analyzing
all
the
source
code
which
is
developed
today.
So
it
was
urgent
to
take
action,
and
this
is
the
reason
why,
a
few
years
ago
we
started
the
project
the
code
name
is
software
heritage
associates
our
heritage.
The
mission
is
very,
at
the
same
time,
humble
and
ambitious.
F
We
are
going
to
actually
look
out
there
to
collect,
preserve
and
maybe
make
available
all
the
source
code
of
all
software
written
evaluated,
only
the
source
code,
all
the
source
code
of
the
software
operator.
It
is
not
just
looking
at
the
path;
it
is,
of
course,
collecting
the
history
and
the
heritage
of
the
software.
That
has
changed
the
other
world,
but
it
is
also
building
an
incredible
telescope
to
observe
the
galaxy
of
the
evolution
of
software
across
the
world.
How
would
you
do
something
like
this
in
a
long-term
perspective?
F
Well,
it
requires
some
thought.
You
know
I'm
old
enough.
I
made
a
lot
of
mistake
in
my
life,
so
I'm
trying
not
to
make
the
same.
At
least
I
mean
to
variate
a
bit
to
beat
the
pleasure
so,
first
of
all
think.
First,
we
need
more
people
sharing
the
vision
you
need
an
international
universal
initiative.
That's
the
reason
we
have
a
partnership
with
unesco
and
many
other
people
in
organizations
supporting
our
work.
This
is
important
to
raise
awareness
of
the
importance
of
this
mission
here.
F
On
the
other
side,
this
of
course
costs
money
and
they
do
not
want
to
depend
on
a
single
donor
or
a
single
organization
or
a
single
structure.
That's
the
reason
we
have
a
bunch
of
different
supporters,
starting
from
a
national
research
infrastructure
in
france,
which
is
called
in
riyadh.
Many
of
you
know
it
and
then
many
many
different
companies.
You
know
there
is
a
lot
of
white
space
there.
F
So
if
anybody
wants
to
join
in,
I
mean
they're
very
they're
going
to
come,
and
how
do
you
do
this
technically
and
why
this
is
more
difficult
than
any
other
archival
process
project?
We
actually
know
well
think
about
it
when
you
want
to
archive
all
the
source
code
or
I
ever
have
written
over
the
planet.
Well
now
you
need
to
go
and
talk
to
guitar.
F
We
are
here,
talk
to
them
happy
to
be
here,
but
you
also
need
to
talk
to
gitlab,
to
bitbucket,
to
npm,
to
pipeline,
to
many
made
to
debian,
to
read
that
many
other
places
where
structural
code
is
actually
available
and
unfortunately
there
is
no
standard.
There
is
no
http
for
the
event
feed
on
these
platforms,
so
we
need
to
build
adapters
for
every
single
one
of
these
and
to
be
able
to
talk
their
language
and
list
the
number
of
projects
which
are
on
each
of
these
platforms.
F
But
that's
not
enough,
then
you
discover
that
a
lot
of
people
use
this
git
today,
but
you
have
tons
of
projects
which
are
under
different
version:
control
system,
subversion,
mercurial,
bazaar
cvs,
you
name
it
and
then
you
have
different
formats
for
packages
etc.
For
a
long-term
perspective,
we
cannot
just
store
a
git
ripper
or
just
store
a
submersion
paper
or
just
store
a
zip
file.
So
we
decided
to
go
one
step
forward
and
we
actually
decided
to
to
translate
the
history
development
of
all
these
kind
of
different
virtual
control
system.
F
So
this
is
actually
building
one
of
the
biggest
public
graph
of
social
expert,
social
production
that
are
available
on
the
planet.
Today
it
is
20
billion
nodes,
280
billion
edges.
I
can
tell
you
for
every
single
file
you
produce
in
how
many
places
it
can
be
found.
Okay,
this
is
an
incredible
source
of
information
and
only
takes
up
400
terabytes,
because
of
course
we
are
de-duplicating
everything.
The
factor
of
compression
is
300.
F
and
then
some
milestone,
and
I
will
just
come
to
the
work
of
today.
I
mean
all
this
started
in
the
summer
2015.
Here
we
could
see
a
picture
of
the
first
server
in
our
data
center.
We
went
public
on
june
2016
with
the
first
sponsors
in
2017.
We've
signed
an
agreement
with
unesco.
Here
we
see
the
former
president
of
french
republic
and
many
others
and
the
former
director
general
of
unesco
on
preservation
of
software.
We
opened
the
archive
last
year
to
the
world,
so
you
can
go
there.
Please
go
there.
F
Archive.Org,
you
will
not
be
lost.
The
interface
is
kind
of
similar
to
the
one
on
github,
but
you
cannot
make
pull
requests.
Okay,
it's
an
archive,
not
a
development
platform.
And
finally,
in
december
last
year
we
opened
up
a
mirror
program
to
have
copies
of
the
archive
under
different
administrative
responsibilities.
All
around
the
planet,
and
in
february
there
was
a
paris
call
by
40
experts
in
the
convene
at
unesco
headquarters,
stating
that
software
source
code
is
important
as
a
part
of
our
cultural
heritage.
F
So
sorry
for
taking
so
much
time.
But
you
see
it's
kind
of
a
passion
of
me
to
be
here
with
you
and
sharing
this.
If
you
want
to
know
more
about
the
project,
just
go
to
www.softrage.org
or
follow
our
twitter
handle
ssw
heritage,
we
see
the
link
to
our
slides
today
and
looking
forward
to
hearing
more
if
you
are
interested
I'll,
be
here
tomorrow.
Also
thanks
a
lot.
Thank
you.
Roberto.
G
So
I'm
wrapping
up
today
and
what
I
thought
I'd
try
to
do
is
give
you
a
little
bit
of
an
idea
of
what
the
geek's
view
is
and
how
wrong
we've
been
about
the
issues
around
long-term
heritage
of
this
stuff,
the
internet
archive.
What
we're
striving
to
be
is
the
library
of
alexandria
for
the
digital
age,
but
the
library
of
alexandria
is
probably
best
known
for
what
burning
right
for
not
being
here
anymore,
or
it
was
the
center
of
learning
for
500
years,
and
it's
just
so
completely
gone
as
best.
G
G
We
can
try
to
figure
out
how
to
get
around
this,
so
I'm
brewster,
kale,
the
founder
and
digital
librarian
of
the
internet
archive-
and
I
thought
I'd-
show
some
of
the
different
attempts
that
are
in
either
my
personal
collection
or
the
internet
archives
collections
of
this
sort
of
long-term
heritage,
thing
and
sort
of
how
is
it
going?
This
is.
This
is
cool.
G
This
is
the
hayakamani
durrani
scor
scroll,
so
japan,
the
empress
shitoku
in
the
year
764,
a.d,
764.,
a.d,
was
became
buddhist
was
really
into
it,
really
wanted
the
sutras
to
last
a
long
time.
So
she
set
out
the
challenge
to
make
one
million
copies
of
the
sutras
and
it's
a
little
scroll
on
the
bottom,
and
that
that
pagoda
that
lathed
pagoda
to
use
about
this
tall
is
is
a
cover.
G
It's
it's
a
it's
a
book
cover,
so
they
made
a
million
of
these
things
and
they
put
them
in
on
10
different
temples
around
japan.
Nine
of
those
are
completely
gone,
but
one
is
still
there
and
they
have
a
ton
of
them
and
they
came
on
hard
times
at
the
end
of
the
1800s,
and
so
they
sold
off
about
3,
000
of
them
and
that's
one
of
them,
which
I'm
kind
of
happy.
G
G
It's
a
picture
of
microfilm
right,
okay!
So
it's
a
little!
You
know
this
was
during
the
20th
century,
the
idea
of
taking
pictures
of
pages
and
having
these
sort
of
funky
readers
to
see
these
old
things
was
kind
of
an
idea
of
universal
access
to
all
knowledge
and
the
medium
was
the
last
500
years
and
actually
it
it
can
and
unless
it's
thrown
out.
So
this
is
a
a
canister
of
print
masters
of
the
periodicals
in
the
united
states.
G
G
This
is
a
the
vault
and
if
you
look
really
carefully
in
the
center
there's
me
I'm
in
in
the
in
the
center
of
that
thing,
and
that's
just
one
row
of
one
floor
of
the
vault
which
went
from
university
microfilms
to
bought
by
xerox,
and
then
it
was
bought
by
pitney
bowles
and
then
it
became
proquest
and
they
spun
it
off
to
na
publishing
which
went
bankrupt
and
it
went
into
a
bank
and
was
going
to
be
thrown
away.
So
we
bought
it.
G
So
the
moral
astoria
here
is
yes,
this
medium
could
have
lasted
for
500
years,
if
you
don't
throw
it
out,
and
so
that's
a
kind
of
a
a
lesson
towards
how
to
do
continuity.
This
is
a
picture
of
the
physical
archive
of
the
internet
archive
in
richmond,
california,
and
it's
basically
in
indoors.
It's
it's
books
or
records,
or
or
whatever
in
boxes,
on
pallets,
surrounded
by
protected
by
boxes
protected
by
shipping
containers
protecting
by
a
warehouse
protected
by
non-profits.
G
G
G
Let's
all
talk
about
tape,
robots,
okay,
so
they
suck,
but
they
were
able
to
get
us
to
20
terabytes
of
collecting
the
world
wide
web
to
become
the
wayback
machine,
starting
in
1996
and
the
ones
behind
us.
We
were
dreaming
of
having
a
terabyte
of
spinning
hard
drives
on
these
sun
microcomputer
era,
things
and
and
other
types
of
of
computers
to
be
able
to
start
to
do
books,
music,
video
web
pages
software
and
we've
been
working
away
at
it.
G
This
is
a
copy
of
the
when
the
library
of
alexandria
in
egypt,
they
made
a
new
one
of
these
libraries.
It
came,
they
were
inaugurated
in
the
year
2002.
We
said,
let's
give
them
a
copy
of
everything,
and
so
we
did
and
it
took
up
most
of
an
airplane
and
it's
these
computers
behind
it
and
they
re-architected
the
first
floor
and
they
were
celebrating
this
digital
collection,
which
is
really
tremendous.
G
G
So
I
can
answer
how
big
is
the
web,
but
you
have
to
ask
me
how
big
is
the
web?
The
web
is
20
feet
by
8
feet
by
eight
feet
and,
and
it
weighs
26
000
pounds,
and
that
makes
it
up
80
micrograms,
a
hit
which
I
think
is
kind
of
fun.
So
that
was
the
era
then,
and
actually
I
think
it's
about
that
same
size
now
given
hard
drives,
are
getting
larger
but
we're
off
of
moore's
law.
So
it's
a
little
bit
of
a
problem.
This
is
the
picture
of
the
current
petaboxes.
G
So
this
is
what
we
currently
put
the
internet
archive.
This
is
a
cut.
This
is
about
seven
and
a
half
petabytes
of
the
primary
copy
of
the
internet
archive
it's
in
a
church
we
bought
in
richmond.
California,
please
visit,
and
the
idea
is
to
try
to
have
these.
This
is
the
primary
part
of
the
primary
copy
of
about
50
petabytes
of
the
internet
archive
it
costs
about
1
500
per
terabyte,
to
try
to
keep
it
a
terabyte
alive
forever
right.
G
So
it
and
that's
sort
of
assuming
still
an
exponential,
and
we
really
need
some
help
out
there
guys
to
keep
going
on
the
decreasing
curve
so
that,
basically,
it's
a
finite
amount
of
money
to
keep
data
alive
and
going.
But
this
is
sort
of
our
current
approach
and
we
have
not
just
the
primary
copy.
You
have
other
copies
and
then
alexandria,
egypt,
amsterdam,
canada,
to
be
able
to
have
other
copies
or
partial
copies
of
this
data
in
in
different
places.
G
But
all
of
this
I've
what
I've
learned
is
really
the
way
to
make
things
live.
Is
these
things,
people
and
people
that
care
that
really,
at
the
end
of
the
day,
I
don't
think
it's
going
to
be
a
hardware
hack.
I
think
it's
going
to
be
people
that
care
about
the
stuff-
and
this
is
a
picture
of
the
internet
archive
staff,
and
these
are
the
folks
that
are
are
caring
at
the
internet
archive.
G
What
is
really
amazing
to
me
about
this
whole
thing
is
github
cared
github
cared
that
they
said.
We
think
all
of
the
work
that's
going
on
all
the
software
that's
being
built
on
top
of
it
is
important
to
keep
around
forever
and
when
john
evans
came
came
to
the
internet
archive
and
said,
would
you
be
interested
in
working
with
github
to
go
and
make
sure
that
this
is
permanently
available
in
a
different
way
than
what
they
were
doing
with
the
others?
We
were
just
overjoyed
right.
G
G
G
So
what
we're
trying
to
do
is
archive
the
whole
damn
thing
and
we
think
it's
about
250
terabytes.
We
want
it
to
not
just
be
a
sort
of
a
you
know:
you're
going
to
have
these
sort
of
in
the
vaults
sort
of
approach
we're
going
to
try
to
make
it
live,
we're
going
to
try
to
make
it
so
that
it's
still
there
you
can
check
in
and
out
you
can.
You
can
talk
the
the
github
protocols,
but
also
be
able
to
be
used
by
researchers
in
new
and
different
ways.